A Multi-Scale CRNN Model for Chinese Papery Medical Document Recognition

Paper-based medical documents are still widely used in many countries, while the contents within are difficult for patients to store and manage. In contrast, electronic medical documents not only help solve these problems, but also promote the development of telemedicine and medical big data. Thus, how to transform traditional printed medical documents into electronic ones becomes a key issue. It is worth noting that recognizing Chinese medical document in image form is a challenging task, as there are a variety of characters and symbols, including Greek alphabets, mathematical symbols and so on. The structure of Chinese characters is also often intricate. At present, the popular Optical Character Recognition methods are designed for single-scale characters, which tend to have poor performance in those complex scenarios. Based on Convolutional Recurrent Neural Network (CRNN), this paper proposes a multi-scale architecture to recognize multi-lingual characters. To verify the effectiveness, the model is trained on a synthetic dataset and evaluated on a real Chinese medical document dataset. The experimental results demonstrate that the proposed method achieves substantial improvement over the recent methods.